# script for topic 7p
#
# to work with percentiles we need a significantly
# large set of values.
source("../gnrnd5.R")
gnrnd5(41583023104, 31000444)
L1
L2 <- sort(L1)
L2
# to find the 41st percentile, find the index
# that is at 0.41*(number of items in the data)
0.41*232
# then look round up and look at that item
L2[ 96 ]
# to find the 95th percentile, find the index
# that is at 0.95*(number of items in the data)
0.95*232
# then look round up and look at that item
L2[ 221 ]
# to find the 43rd percentile, find the index
# that is at 0.43*(number of items in the data)
0.43*232
# then look round up and look at that item
L2[ 100 ]
# R provides a function to get percentiles from our
# data, without even sorting the data. This is the
# quantile function.
# Re-do the problems
quantile(L1, 0.41)
quantile(L1, 0.95)
quantile(L1, 0.43) # this gives us a different answer
# let us see if we can get some info on this
?quantile
# and we can give quantile() a whole list of
# percentiles to find
quantile( L1, c(0.20, 0.33, 0.14, 0.78))
# or even a sequence of values
quantile( L1, seq(0.60,0.95,0.05))
# If we want to know what percentile is the value
# 487 we find 487 in the sorted data and then get
# its position
which( L2==487)
# then find the percent that index is of the
# total size of the data and then round down,
# but we never go over 99%
213/232*100
#